Facebook made international headlines just recently for what looked like stumbling into an artificial learning technology that could kick start the onset of Skynet.
Skynet being an AI-powered advance robotics character in the Terminator movies sequels that had a philosophical revelation that humans are irreparably bad that eliminating them off the face of the Earth and replacing them with robots will make the world a better place.
Facebook in its beta development program has two artificial intelligence (AI) bots talk to each other. The end goal of this test was to put the negotiation game of the two bots in test and have them reach a deal, in a manner akin to human discussing with one another and coming to an agreement.
Things obviously did not go as planned, as the two machines quickly started creating shorthand and communicating to one another in a manner not readily coherent to the typical human. The two bots seemed like they had quickly invented a new language only understood by them, and no human.
Thanks to all the horrific sci-fi movies Hollywood has served us through the years, the strange ‘language’ by the two bots struck a self-preservation nerve as a human species against a possible robot apocalypse.
The reality of the situation
As interesting (in a weird way) as Facebook creating an AI technology that went rogue and is now trying to conquer the human world, the truth is rather disappointing. The fact is, Facebook made a human error in not restricting the two bots conversation to only using human-coherent language.
As it was, the two bots machine learning algorithm gave them the freedom to choose whichever language they please. They then started creating short hands that were understandable to them as bots, but not to humans. That is as far as the problem goes, and has nothing at all to do with assumptions made by many people that Facebook had stumbled into a technology that will kick start the onset of the robots apocalypse. The fix to this problem is Facebook restricting the bots to conversing in language readily coherent to the typical human.
For more details on this, click here.